Re: Strawman Promises consensus position, based on Thursday's telechat

On 06/10/2014 11:44 AM, Adam Roach wrote:
> On 10/6/14 10:39, cowwoc wrote:
>> On 06/10/2014 11:36 AM, Adam Roach wrote:
>>> On 10/6/14 10:14, cowwoc wrote:
>>> >> navigator.getUserMedia = navigator.getUserMedia || 
>>> navigator.webkitGetUserMedia || navigator.mozGetUserMedia;
>>> ...
>>>> The argument goes that we can provide a shim for Promises -> 
>>>> Callbacks in the same way that you handle prefixes below.
>>>
>>> There's a broad difference between "can be done" and "is already 
>>> done, nearly universally."
>>
>> I don't understand your reasoning. We don't have to wait for the shim 
>> to exist to decide on whether it's worthwhile going down that route. 
>> Are you implying that there would be major difficulties in 
>> implementing this?
>
> No -- I'm saying that there is a significant corpus of code on the 
> web, some as old as 18 months or more, that already uses the construct 
> Chris describes. If Chrome and Firefox pull the prefix off of gUM 
> today, none of that breaks.
>
> On the other hand, approximately zero percent of this code has the 
> shim you mention. It's a materially different argument that code *can 
> be made to work* with a change than it is to say that code *will 
> seamlessly work* with a change.
>
> Unprefixing will seamlessly work.
>
> The same cannot be said for the shim you mention because it's not in 
> deployed code.

Okay, but unprefixing does not happen on its own. A developer needs to 
manually change the application code to make it happen.

We are asking developers to add 3 lines of code at the same time as 
prefixing:

Navigator.prototype.webkitGetUserMedia = function(constraints, success, 
failure) {
  navigator.mediaDevices.getUserMedia(constraints).then(success, failure);
};

Is that really a big deal?

Gili